CLDR-17686 change enum.valueOf to enum.forString in a couple of places

[srl295]

• these were causing odd compilation errors on one environment

CLDR-17686

• This PR completes the ticket.

ALLOW_MANY_COMMITS=true

[srl295]

Deleting maven cache seemed to make this error go away.
So, not sure of the cause.

[btangmu]

I guess this returns null if it catches the exception?

        public static PageId forString(String name) {
            try {
                return PageIdNames.forString(name);
            } catch (Exception e) {
                throw new ICUException("No PageId for " + name, e);
            }
        }

[btangmu]

that's a feature of Java I wasn't familiar with

[srl295]

no, that would throw.

[btangmu]

Oh, right; I misread it. It catches IllegalArgumentException then throws ICUException.

So the removal of catch (Throwable t) means that searchPathheader may throw ICUException where before it wouldn't have. That goes beyond merely changing valueOf to forString. Here q comes from an http request, so it could be anything... Line 765 of SurveyAjax will catch it:

        } catch (Throwable e) {
            SurveyLog.logException(
                    logger, e, "Error in SurveyAjax?what=" + what + ": " + e.getMessage());
            sendError(out, e);
        }

I guess that's OK, though shouldn't searchPathheader then be declared to throw ICUException? It still silently catches other exceptions further below; it just seems a little arbitrary...

[btangmu]

Oh, right; I misread it. It catches IllegalArgumentException then throws ICUException.

So the removal of catch (Throwable t) means that searchPathheader may throw ICUException where before it wouldn't have. That goes beyond merely changing valueOf to forString. Here q comes from an http request, so it could be anything... Line 765 of SurveyAjax will catch it:

        } catch (Throwable e) {
            SurveyLog.logException(
                    logger, e, "Error in SurveyAjax?what=" + what + ": " + e.getMessage());
            sendError(out, e);
        }

I guess that's OK, though shouldn't searchPathheader then be declared to throw ICUException? It still silently catches other exceptions further below; it just seems a little arbitrary...

[macchiati]

Please take a pass through PRs to make sure that they are merged before we branch (if there are other merges on the ticket)